OWASP Red Teaming Solutions Taxonomy Explained | What is the are red, blue and purple teaming

Every AI security vendor claims "red teaming." Without a shared capability grid, you cannot compare solutions or find lifecycle gaps. OWASP's GenAI Security Project just published a taxonomy for the AI and Agentic Red Teaming Solutions Landscape, and it answers a different question than the Top 10: → OWASP Top 10 / ASI: what risks exist → This taxonomy: what capabilities your tooling must cover Three team colors: 🔴 Red: adversarial suites, multi-turn attacks, autonomous red bots 🔵 Blue: guardrail conformance, LLM/agent firewall, runtime AI-SPM 🟣 Purple: one-click purple runs, alert correlation, closed-loop coaching Most teams over-invest in deploy-time pen tests and under-invest in Test & Evaluation, Operate, and Govern. New video: how to use the taxonomy to evaluate AI security tooling. OWASP resource: https://genai.owasp.org/resource/ai-security-solutions-landscape-for-ai-and-agentic-red-teaming-q2-2026/